eComment. Everything changes even statistics: It is time to use bootstrapped confidence intervals?

0 Authors: Ugur Kucuk, Hilan Olgun Kucuk, Kadir Hakan Cansiz and Onur Durmaz Van Army District Hospital , Van, Turkey doi: 10.1093/icvts/ivu135 The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved - We have read the well-written article by Lio et al. with great interest [1]. The authors compared mitral valve repair and mitral valve replacement in patients with ischaemic mitral regurgitation and depressed ejection fraction. They obtained invaluable scientific data from the study. These results can be represented with a stronger level of evidence by utilizing a robust statistical method called bootstrapping. A confidence interval (CI) gives an estimated range of values which is likely to include an unknown population parameter, the estimated range being calculated from a given set of sample data. Confidence intervals are needed because there is variation in nature; nearly all information gained from humans varies to a greater or lesser extent. There are two important factors that affect the width of a CI: the sample size and the amount of variation in the population. Classically, CIs are calculated with formulas developed on the assumptions of normality and the central limit theorem which were developed when there were no computers, and analytical methods were needed in the absence of computational power. How do we know how much sample statistics vary, if we only have one sample? The answer lies in the term bootstrapping. In essence you use the sample data to take large numbers of random samples and examine the distribution of these samples. You can do it by re-using the data from your one actual study over and over again. The term bootstrapping is an allusion to the expression pulling oneself up by ones bootstraps, in this case using the sample data as a population from which repeated samples are drawn. Over the years, the bootstrap procedure has become an accepted way to get reliable estimates of standard errors (SE) and confidence intervals for almost anything you can calculate from your data [2]. Nowadays bootstrapping is often considered the gold standard method to determine SEs and CIs. Bootstrap techniques are heavily dependent upon computer calculations. As a widely used programme for statistical analysis in medicine, SPSS 18 and newer versions afford bootstrap methods for standard use. Bootstrap based approaches for statistical estimation and determination of the properties of the estimator are being increasingly realized in modern methods of data analysis. As a result it is time to revise our statistical habits. Conflict of interest: none declared. (...truncated)